Use sha instead of date to match patches with deploys

calculate_four_metrics.py

@@ -26,7 +26,7 @@
 import docopt
 
 from git_metrics_release_lead_time import commit_author_time_tag_author_time_and_from_to_tag_name, \
-    fetch_tags_and_author_dates, fetch_tags_and_commit_dates
+    fetch_tags_and_author_dates, fetch_tags_and_sha
 from process import mk_run
 from recovery_time import Deployment, find_is_patch, find_outages
 
@@ -69,24 +69,17 @@ def main():
 
 def calculate_MTTR(path_to_git_repo, deploy_pattern, patch_pattern, start_date):
     run = mk_run(path_to_git_repo)
+    match_deploy = partial(fnmatch, pat=deploy_pattern)
     deploy_tags_author_date = fetch_tags_and_author_dates(
         run,
-        partial(fnmatch, pat=deploy_pattern),
+        match_deploy,
         start_date,
     )
-    deploy_tags_commit_date = dict(fetch_tags_and_commit_dates(
-        run,
-        partial(fnmatch, pat=deploy_pattern),
-        start_date,
-    ))
+    deploy_tags_commit_date = dict(fetch_tags_and_sha(run, match_deploy))
     patch_dates = set(
         date
         for _tag, date
-        in fetch_tags_and_commit_dates(
-            run,
-            partial(fnmatch, pat=patch_pattern),
-            start_date,
-        )
+        in fetch_tags_and_sha(run, partial(fnmatch, pat=patch_pattern))
     )
     deployments = []
     for deploy_tag, deploy_date in deploy_tags_author_date:

git_metrics_release_lead_time.py

@@ -14,11 +14,6 @@
     sort='taggerdate'
 )
 
-TAGS_WITH_COMMIT_DATE_CMD = for_each_ref(
-    'refs/tags/**',
-    format='%(refname:short) %(*authordate:unix)',
-)
-
 TAGS_WITH_COMMIT_SHA_CMD = for_each_ref(
     'refs/tags/**',
     format='%(refname:short) %(*objectname)',
@@ -31,12 +26,6 @@ def tags_with_author_date(run) -> Iterable[Tuple[str, int]]:
     return parse_tags_with_date(stdout)
 
 
-def tags_with_commit_date(run) -> Iterable[Tuple[str, int]]:
-    proc = run(TAGS_WITH_COMMIT_DATE_CMD)
-    stdout = proc_to_stdout(proc)
-    return parse_tags_with_date(stdout)
-
-
 def parse_tags_with_date(lines: Iterable[str]) -> Iterable[Tuple[str, int]]:
     return ((tag_and_maybe_date[0], int(tag_and_maybe_date[1])) for tag_and_maybe_date in columns(lines) if len(tag_and_maybe_date) > 1)
 
@@ -77,27 +66,17 @@ def fetch_tags_and_author_dates(run, match_tag, earliest_date=0):
     return filtered_on_tags_date
 
 
-def fetch_tags_and_sha(run):
+def fetch_tags_and_sha(run, match_tag):
     proc = run(TAGS_WITH_COMMIT_SHA_CMD)
     stdout = proc_to_stdout(proc)
     return (
         (tag_and_maybe_sha[0], tag_and_maybe_sha[1])
         for tag_and_maybe_sha
         in columns(stdout)
-        if len(tag_and_maybe_sha) > 1
+        if len(tag_and_maybe_sha) > 1 and match_tag(tag_and_maybe_sha[0])
     )
 
 
-
-def fetch_tags_and_commit_dates(run, match_tag, earliest_date=0):
-    tags_and_date = tags_with_commit_date(run)
-    filtered_on_tags_date = filter(
-        lambda p: match_tag(p[0]) and p[1] > earliest_date,
-        tags_and_date
-    )
-    return filtered_on_tags_date
-
-
 def plot_release_lead_time_metrics(data):
     import matplotlib.pyplot as plt
     from pandas import DataFrame

test_release_lead_time.py

@@ -33,7 +33,7 @@ def test_fetch_tags_and_sha():
     result = fetch_tags_and_sha(lambda _: stdout([
         "annotated-tag sha",
         "lightweight-tag"
-    ]))
+    ]), lambda _: True)
     assert list(result) == [
         ("annotated-tag", "sha")
     ]